To provide 'cbc' executable path.

[fgunyel]

settings = dict(solver='cbc', solve_kw={ 'tee': False, # print solver output }, solver_cmdline_options={

'allowableGap': 1e-5, # (absolute gap) default: 1e-10

                # 'ratioGap': 0.2,  # (0.2 = 20% gap) default: 0
                # 'seconds': 60 * 1,  # (maximum runtime) default: 1e+100
            },
            )

network.optimize_investment(invest_options=invest_opt, **settings)

gives an error:

ApplicationError: No executable found for solver 'cbc'

Altough cbc is installed. I want to be able to give its path. Is this possible?

[joroeder]

If you are using Windows, you might still need to add the installation path to your path variable. Please see the CBC installation instructions, or follow https://oemof-solph.readthedocs.io/en/latest/readme.html#installing-a-solver

[fgunyel]

Yes, that is exactly what I am looking for. The problem is that I am not able to specify the path for the "settings" since it does only take the solver name as input, but not where it is located. It must be something right:

settings = dict(solver='cbc', path=e.g. C:/Users/Somebody/my_program, .....)

The cbc is already succesfully installed in the PC.

[joroeder]

You need to add the solver to your PATH variable in Windows before. This has nothing to do with DHNx and the attributes of the optimisation method. Please see https://www.computerhope.com/issues/ch000549.htm

[fgunyel]

Thank you Johannes. I am working on Google Colab. I will try to find a way to define path variables there if I can. If I succeed, I put an update here.

[fgunyel]

!apt-get install -y -qq coinor-cbc import os !export cbc='/usr/bin/cbc' os.environ['cbc'] ='/usr/bin/cbc

The above code in Google Colab solves the problem, prior to calling cbc. Great!